warming up your workspaceLoading
Systematically diagnose network performance problems including TCP retransmissions, conntrack table overflow, MTU/MSS mismatches, and SYN floods using ss, tcpdump, and kernel counters.
## Problem
You are an SRE at a company running thousands of Linux servers. Users report intermittent latency spikes and connection failures to your services. Your monitoring shows elevated TCP retransmissions, occasional connection timeouts, and throughput degradation on certain paths. Describe your systematic approach to diagnosing these network performance issues, covering the tools, kernel metrics, and common root causes you would investigate.
Sign up to access the full problem
Design canvas, rubric, hints, and model solutions.